You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@hive.apache.org by "jhuchuan (Jira)" <ji...@apache.org> on 2022/09/27 14:28:00 UTC
[jira] [Updated] (HIVE-26570) Incorrect results on sum(nvl(col,0)) when vectorization is ON
[ https://issues.apache.org/jira/browse/HIVE-26570?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
jhuchuan updated HIVE-26570:
----------------------------
Description:
1、
create table testdb.lc_appl
(
loan_no string,
fee_amt decimal(16,2)
)
clustered by (loan_no)
into 5 buckets
stored as orc
tblproperties('transactional'='true');
2、
insert into testdb.lc_appl
values ('a',12.12)
insert into testdb.lc_appl
values ('b',13.13)
set hive.vectorized.execution.enabled=false;
select loan_no ,sum(fee_amt),sum(nvl(fee_amt,0))
from testdb.lc_appl
group by loan_no
--right result
a 12.12 12.12
b 13.13 13.13
set hive.vectorized.execution.enabled=true;
select loan_no ,sum(fee_amt),sum(nvl(fee_amt,0))
from testdb.lc_appl
group by loan_no
--wrong result
a 12.12 0
b 13.13 0
was:
1、
create table testdb.lc_appl
(
loan_no string,
fee_amt decimal(16,2)
)
clustered by (loan_no)
into 5 buckets
stored as orc
tblproperties('transactional'='true');
2、
insert into testdb.lc_appl
values ('a',12.12)
insert into testdb.lc_appl
values ('b',13.13)
set hive.vectorized.execution.enabled=false;
select loan_no ,sum(fee_amt),sum(nvl(fee_amt,0))
from testdb.lc_appl
group by loan_no
--result
a 12.12 12.12
b 13.13 13.13
set hive.vectorized.execution.enabled=true;
select loan_no ,sum(fee_amt),sum(nvl(fee_amt,0))
from testdb.lc_appl
group by loan_no
--result
a 12.12 0
b 13.13 0
> Incorrect results on sum(nvl(col,0)) when vectorization is ON
> -------------------------------------------------------------
>
> Key: HIVE-26570
> URL: https://issues.apache.org/jira/browse/HIVE-26570
> Project: Hive
> Issue Type: Bug
> Components: Vectorization
> Affects Versions: 2.1.1
> Reporter: jhuchuan
> Priority: Major
> Fix For: All Versions
>
>
> 1、
> create table testdb.lc_appl
> (
> loan_no string,
> fee_amt decimal(16,2)
> )
> clustered by (loan_no)
> into 5 buckets
> stored as orc
> tblproperties('transactional'='true');
>
> 2、
> insert into testdb.lc_appl
> values ('a',12.12)
> insert into testdb.lc_appl
> values ('b',13.13)
>
> set hive.vectorized.execution.enabled=false;
> select loan_no ,sum(fee_amt),sum(nvl(fee_amt,0))
> from testdb.lc_appl
> group by loan_no
> --right result
> a 12.12 12.12
> b 13.13 13.13
>
>
> set hive.vectorized.execution.enabled=true;
> select loan_no ,sum(fee_amt),sum(nvl(fee_amt,0))
> from testdb.lc_appl
> group by loan_no
> --wrong result
> a 12.12 0
> b 13.13 0
>
>
--
This message was sent by Atlassian Jira
(v8.20.10#820010)